Bhatt Bhumi J, Ghosh Sunita, Mazurak Vera, Brun Aurélien Q, Bathe Oliver, Baracos Vickie E, Damaraju Sambasivarao
Department of Oncology, Faculty of Medicine and Dentistry, University of Alberta, Edmonton, Alberta, Canada.
Department of Laboratory Medicine and Pathology, Faculty of Medicine and Dentistry, University of Alberta, Edmonton, Alberta, Canada.
Nature. 2025 Sep 10. doi: 10.1038/s41586-025-09502-0.
Cancer-associated muscle wasting is associated with poor clinical outcomes, but its underlying biology is largely uncharted in humans. Unbiased analysis of the RNAome (coding and non-coding RNAs) with unsupervised clustering using integrative non-negative matrix factorization provides a means of identifying distinct molecular subtypes and was applied here to muscle of patients with colorectal or pancreatic cancer. Rectus abdominis biopsies from 84 patients were profiled using high-throughput next-generation sequencing. Integrative non-negative matrix factorization with stringent quality metrics for clustering identified two highly coherent molecular subtypes within muscle of patients with cancer. Patients with subtype 1 (versus subtype 2) showed clinical manifestations of cachexia: high-grade weight loss, low muscle mass, atrophy of type IIA and type IIX muscle fibres, and reduced survival. On the basis of differential expression between the subtypes, we identified biological processes that may contribute to cancer-associated loss of muscle mass and function, including altered posttranscriptional regulation and perturbation of neuronal systems; cytokine storm and cellular immune response; pathways related to extracellular matrix; and metabolic abnormalities spanning xenobiotic metabolism, haemostasis, signal transduction, embryonic and/or pluripotent stem cells, and amino acid metabolism. Differential expression between subtypes indicated the involvement of multiple intertwined higher-order gene regulatory networks, suggesting that network interactions of (hub) long non-coding RNAs, microRNAs and mRNAs could represent targets for future research.
癌症相关的肌肉萎缩与不良临床结局相关，但其潜在生物学机制在人类中很大程度上仍未被探索。使用整合非负矩阵分解进行无监督聚类对RNA组（编码和非编码RNA）进行无偏分析，提供了一种识别不同分子亚型的方法，并在此应用于结直肠癌或胰腺癌患者的肌肉。对84例患者的腹直肌活检组织进行了高通量二代测序分析。使用严格的聚类质量指标进行整合非负矩阵分解，在癌症患者的肌肉中识别出两种高度一致的分子亚型。1型患者（相对于2型）表现出恶病质的临床表现：严重体重减轻、低肌肉量、IIA型和IIX型肌纤维萎缩以及生存率降低。基于亚型之间的差异表达，我们确定了可能导致癌症相关肌肉量和功能丧失的生物学过程，包括转录后调控改变和神经系统紊乱；细胞因子风暴和细胞免疫反应；与细胞外基质相关的通路；以及涉及异源生物代谢、止血、信号转导、胚胎和/或多能干细胞以及氨基酸代谢的代谢异常。亚型之间的差异表达表明多个相互交织的高阶基因调控网络的参与，提示（枢纽）长链非编码RNA、微小RNA和信使RNA的网络相互作用可能代表未来研究的靶点。
